
- Outline-from-series-of-triangles
- 08-16-2007
![]() Re: Outline from series of triangles
| Dave Eberly | 08-16-2007 |
![]() ![]() Re: Outline from series of triangles
| Brian Cryer | 08-17-2007 |
If you were Registered and logged in, you could reply and use other advanced thread options
We have an in-house application that generates a series of overlapping
triangles (thousands) and then traces round the edge of these (and displays
the results as an overlay on a map). The tracing algorithm isn't
particularly fast and was developed by someone who left the company years
ago. I was wondering if there are established algorithms for this type of
thing - because I would like to investigate alternatives to see if I can
speed up what we currently do.
TIA.
triangles (thousands) and then traces round the edge of these (and displays
the results as an overlay on a map). The tracing algorithm isn't
particularly fast and was developed by someone who left the company years
ago. I was wondering if there are established algorithms for this type of
thing - because I would like to investigate alternatives to see if I can
speed up what we currently do.
TIA.
http://www.geometrictools.com/SampleFoundation/MeshEnvelope/MeshEnvelope.html
The application is currently configured to use exact arithmetic for robust
calculations, but it may be easily switched to using fixed-precision
floating-point arithmetic.
The bottleneck in such an algorithm is determining all the edge-edge
intersections of the triangles. My implementation uses a sweep
method to keep this cost to a minimum (file CreateEnvelope.cpp,
function UpdateAllEdges).
--
Dave Eberly
http://www.geometrictools.com
>> We have an in-house application that generates a series of overlapping
>> triangles (thousands) and then traces round the edge of these (and
>> displays the results as an overlay on a map). The tracing algorithm isn't
>> particularly fast and was developed by someone who left the company years
>> ago. I was wondering if there are established algorithms for this type of
>> thing - because I would like to investigate alternatives to see if I can
>> speed up what we currently do.
>> triangles (thousands) and then traces round the edge of these (and
>> displays the results as an overlay on a map). The tracing algorithm isn't
>> particularly fast and was developed by someone who left the company years
>> ago. I was wondering if there are established algorithms for this type of
>> thing - because I would like to investigate alternatives to see if I can
>> speed up what we currently do.
> http://www.geometrictools.com/SampleFoundation/MeshEnvelope/MeshEnvelope.html
> The application is currently configured to use exact arithmetic for robust
> calculations, but it may be easily switched to using fixed-precision
> floating-point arithmetic.
> The bottleneck in such an algorithm is determining all the edge-edge
> intersections of the triangles. My implementation uses a sweep
> method to keep this cost to a minimum (file CreateEnvelope.cpp,
> function UpdateAllEdges).
> The application is currently configured to use exact arithmetic for robust
> calculations, but it may be easily switched to using fixed-precision
> floating-point arithmetic.
> The bottleneck in such an algorithm is determining all the edge-edge
> intersections of the triangles. My implementation uses a sweep
> method to keep this cost to a minimum (file CreateEnvelope.cpp,
> function UpdateAllEdges).
Thank you Dave. I'm off for a few days but once I'm back I'll take a long
hard look at this.
- A ceramic filter followed by a series of surface acoustic wave (SAW) filters
- Satellite Navigation
- 2011-11-29
- Data for eTrex series
- Garmin GPS
- 2010-04-29
- eTrex H series v previous models
- Garmin GPS
- 2009-11-18
- Nuvi 500 series and Snowmobiling
- Garmin GPS
- 2009-09-05
- New series of Acer GPS navigators
- Magellan GPS
- 2009-04-23






> triangles (thousands) and then traces round the edge of these (and
> displays the results as an overlay on a map). The tracing algorithm isn't
> particularly fast and was developed by someone who left the company years
> ago. I was wondering if there are established algorithms for this type of
> thing - because I would like to investigate alternatives to see if I can
> speed up what we currently do.